Naive Bayesian prediction of bleeding after heart by-pass surgery

Abstract

Excessive post-operative bleeding occurs in approximately one out of eight patients who undergo heart bypass surgery. Earlier workers have identified laboratory parameters that are correlated with post-operative blood loss but these correlations are not strong enough to be clinically useful. This paper describes a predictor that combines several of these parameters using Naive Bayesian Reasoning, to produce a clinically useful predictor of blood loss.